Neural processor in artificial intelligence advancement

Mitra, Manu

doi:10.32629/jai.v1i1.13

Cited by 3 publications

(3 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Em seguida, solicita-se à rede que resolva um problema. Ela tenta resolver repetidamente, sempre fortalecendo as conexões que levam ao sucesso e diminuindo as que levam ao fracasso (KOCEJKO, 2018;MITRA, 2018) Quanto ao procedimento técnico, a pesquisa pode ser classificada em bibliográfica. A pesquisa bibliográfica é desenvolvida a partir de material já elaborado (GIL, 2008) e implica no estudo de artigos, teses, livros e outras publicações usualmente disponibilizadas por editoras e indexadas (WAZLAWICK, 2014).…”

Section: Rede Neuralunclassified

Absorção das tarefas de processamento de Linguagem Natural (NLP) pela Ciência da Informação (CI)

Falcão

Lopes

Souza

2021

View full text Add to dashboard Cite

Um dos recentes desafios da abordagem denominada Big Data tem sido extrair informações relevantes de grandes quantidades de dados não estruturados, como por exemplo de textos escritos em diversos idiomas. A principal abordagem de análise de texto e linguagem por meio computacional é chamada de Processamento de Linguagem Natural (Natural Language Processing - NLP, na sigla em inglês). Identificar como as áreas do conhecimento estão utilizando as evoluções em seus domínios, no que tange à NLP, especialmente a Ciência da Informação, por fornecer os principais conceitos de tratamento de dados, informações e conhecimento, é o cerne desse estudo. Este foi estruturado tendo como base uma Revisão Sistemática da Literatura, entendendo ser essa uma abordagem capaz de fundamentar noções iniciais e, ao mesmo tempo, consistentes para a análise da questão central que motivou esse trabalho. Dentre os resultados encontrados está a pouca utilização dos recursos de NLP pela Ciência da Informação.

show abstract

Section: Rede Neuralunclassified

Absorção das tarefas de processamento de Linguagem Natural (NLP) pela Ciência da Informação (CI)

Falcão

Lopes

Souza

2021

View full text Add to dashboard Cite

show abstract

“…The increase of the working memory load in the motor imagery is often accompanied with the increase of the power of θ wave, while the wave is often not considered in the BCI study. Event-Related Synchronization (ERS) [9] refers to the increase of cortical activity in specific frequency bands (such as and bands), and Event-Related Desynchronization (ERD) [10] refers to the reduction of cortical activity in specific frequency bands (such as and bands). They mainly reflect the changes of EEG amplitude caused by motor imagery.…”

Section: Signal Acquisitionmentioning

confidence: 99%

Research on the Key Technologies of Motor Imagery EEG Signal Based on Deep Learning

Wang

et al. 2020

J Autonom Intell

View full text Add to dashboard Cite

Brain-computer interface (BCI) is an emerging area of research that establishes a connection between the brain and external devices in a completely new way. It provides a new idea about the rehabilitation of brain diseases, human-computer interaction and augmented reality. One of the main problems of implementing BCI is to recognize and classify the motor imagery Electroencephalography(EEG) signals effectively. This paper takes the motor imagery feature data of EEG as the research object to conduct the research of multi-classification method. In this study, we use the Emotiv helmet with 16 biomedical sensors to obtain EEG signal, adopt the fast independent component analysis and the fast Fourier transform to realize signal preprocessing and select the common spatial pattern algorithm to extract the features of the motor imagery EEG signal. In order to improve the accuracy of recognition of EEG signal, a new deep learning network is designed for multi-channel self-acquired EEG data set which is named as min-VGG-LSTMnet in this paper. The network combines Long Short-Term Memory Network with convolutional neural network VGG and achieves the four-class task of the left-hand, right-hand, left-foot and right-foot lifting movements based on motor imagery. The results show that the accuracy of the proposed classification method is at least 8.18% higher than other mainstream deep learning methods.

show abstract

“…By taking as reference the path followed by the speech recognition research, and considering current trends in embedded machine learning models, one could identify some topics that should be addressed in order to make small ASC systems viable. Examples are the research on i) frameworks such as [6] and [13] for on-device training of deep neural networks (DNNs), ii) Convolutional neural networks (CNNs) small enough to be embedded [8], [9], iii) segmentation techniques to reduce the size of existing datasets to allow more efficient model training and evaluation [14], be it for small models [15], or few-shot learning frameworks [16]. Although topic i) is important in applications in which online training is not always possible (e.g., smartphones and domestic assistants which need to work even in the absence of internet connection), one could argue that topics ii) and iii) necessarily impact the feasibility of i).…”

Section: Introductionmentioning

confidence: 99%

On creating small datasets for training embedded acoustic scene classification systems via time-frequency segmentation

Souza¹,

Ferreira²,

Ferreira³

et al. 2021

Anais Do XXXIX Simpósio Brasileiro De Telecomunicações E Processamento De Sinais

View full text Add to dashboard Cite

Acoustic Scene Classification (ASC) systems have great potential to transform existing embedded technologies. However, research on ASC has put little emphasis on solving the existing challenges in embedding ASC systems. In this paper, we focus on one of the problems associated with smaller ASC models: the generation of smaller yet highly informative training datasets. To achieve this goal, we propose to employ the so-called multitaper-reassignment technique to generate high-resolution spectrograms from audio signals. These sharp time-frequency (TF) representations are used as inputs to a splitting method based on TF-related entropy metrics. We show via simulations that the datasets created through the proposed segmentation can successfully be used to train small convolutional neural networks (CNNs), which could be employed in embedded ASC applications.

show abstract

Neural processor in artificial intelligence advancement

Cited by 3 publications

References 1 publication

Absorção das tarefas de processamento de Linguagem Natural (NLP) pela Ciência da Informação (CI)

Absorção das tarefas de processamento de Linguagem Natural (NLP) pela Ciência da Informação (CI)

Research on the Key Technologies of Motor Imagery EEG Signal Based on Deep Learning

On creating small datasets for training embedded acoustic scene classification systems via time-frequency segmentation

Contact Info

Product

Resources

About